old repos: use /pve suffix, which has better compatibility
[pve-docs.git] / qm.adoc
CommitLineData
80c0adcb 1[[chapter_virtual_machines]]
f69cfd23 2ifdef::manvolnum[]
b2f242ab
DM
3qm(1)
4=====
5f09af76
DM
5:pve-toplevel:
6
f69cfd23
DM
7NAME
8----
9
10qm - Qemu/KVM Virtual Machine Manager
11
12
49a5e11c 13SYNOPSIS
f69cfd23
DM
14--------
15
16include::qm.1-synopsis.adoc[]
17
18DESCRIPTION
19-----------
20endif::manvolnum[]
f69cfd23
DM
21ifndef::manvolnum[]
22Qemu/KVM Virtual Machines
23=========================
5f09af76 24:pve-toplevel:
194d2f29 25endif::manvolnum[]
5f09af76 26
c4cba5d7
EK
27// deprecates
28// http://pve.proxmox.com/wiki/Container_and_Full_Virtualization
29// http://pve.proxmox.com/wiki/KVM
30// http://pve.proxmox.com/wiki/Qemu_Server
31
5eba0743 32Qemu (short form for Quick Emulator) is an open source hypervisor that emulates a
c4cba5d7
EK
33physical computer. From the perspective of the host system where Qemu is
34running, Qemu is a user program which has access to a number of local resources
35like partitions, files, network cards which are then passed to an
189d3661 36emulated computer which sees them as if they were real devices.
c4cba5d7
EK
37
38A guest operating system running in the emulated computer accesses these
3a433e9b
OB
39devices, and runs as if it were running on real hardware. For instance, you can pass
40an ISO image as a parameter to Qemu, and the OS running in the emulated computer
41will see a real CD-ROM inserted into a CD drive.
c4cba5d7 42
6fb50457 43Qemu can emulate a great variety of hardware from ARM to Sparc, but {pve} is
c4cba5d7
EK
44only concerned with 32 and 64 bits PC clone emulation, since it represents the
45overwhelming majority of server hardware. The emulation of PC clones is also one
46of the fastest due to the availability of processor extensions which greatly
47speed up Qemu when the emulated architecture is the same as the host
9c63b5d9
EK
48architecture.
49
50NOTE: You may sometimes encounter the term _KVM_ (Kernel-based Virtual Machine).
51It means that Qemu is running with the support of the virtualization processor
3a433e9b
OB
52extensions, via the Linux KVM module. In the context of {pve} _Qemu_ and
53_KVM_ can be used interchangeably, as Qemu in {pve} will always try to load the KVM
9c63b5d9
EK
54module.
55
c4cba5d7
EK
56Qemu inside {pve} runs as a root process, since this is required to access block
57and PCI devices.
58
5eba0743 59
c4cba5d7
EK
60Emulated devices and paravirtualized devices
61--------------------------------------------
62
189d3661 63The PC hardware emulated by Qemu includes a mainboard, network controllers,
3a433e9b 64SCSI, IDE and SATA controllers, serial ports (the complete list can be seen in
189d3661
DC
65the `kvm(1)` man page) all of them emulated in software. All these devices
66are the exact software equivalent of existing hardware devices, and if the OS
67running in the guest has the proper drivers it will use the devices as if it
c4cba5d7
EK
68were running on real hardware. This allows Qemu to runs _unmodified_ operating
69systems.
70
71This however has a performance cost, as running in software what was meant to
72run in hardware involves a lot of extra work for the host CPU. To mitigate this,
73Qemu can present to the guest operating system _paravirtualized devices_, where
74the guest OS recognizes it is running inside Qemu and cooperates with the
75hypervisor.
76
470d4313 77Qemu relies on the virtio virtualization standard, and is thus able to present
189d3661
DC
78paravirtualized virtio devices, which includes a paravirtualized generic disk
79controller, a paravirtualized network card, a paravirtualized serial port,
c4cba5d7
EK
80a paravirtualized SCSI controller, etc ...
81
189d3661
DC
82It is highly recommended to use the virtio devices whenever you can, as they
83provide a big performance improvement. Using the virtio generic disk controller
84versus an emulated IDE controller will double the sequential write throughput,
85as measured with `bonnie++(8)`. Using the virtio network interface can deliver
c4cba5d7 86up to three times the throughput of an emulated Intel E1000 network card, as
189d3661 87measured with `iperf(1)`. footnote:[See this benchmark on the KVM wiki
a55d30db 88https://www.linux-kvm.org/page/Using_VirtIO_NIC]
c4cba5d7 89
5eba0743 90
80c0adcb 91[[qm_virtual_machines_settings]]
5274ad28 92Virtual Machines Settings
c4cba5d7 93-------------------------
80c0adcb 94
c4cba5d7
EK
95Generally speaking {pve} tries to choose sane defaults for virtual machines
96(VM). Make sure you understand the meaning of the settings you change, as it
97could incur a performance slowdown, or putting your data at risk.
98
5eba0743 99
80c0adcb 100[[qm_general_settings]]
c4cba5d7
EK
101General Settings
102~~~~~~~~~~~~~~~~
80c0adcb 103
1ff5e4e8 104[thumbnail="screenshot/gui-create-vm-general.png"]
b16d767f 105
c4cba5d7
EK
106General settings of a VM include
107
108* the *Node* : the physical server on which the VM will run
109* the *VM ID*: a unique number in this {pve} installation used to identify your VM
110* *Name*: a free form text string you can use to describe the VM
111* *Resource Pool*: a logical group of VMs
112
5eba0743 113
80c0adcb 114[[qm_os_settings]]
c4cba5d7
EK
115OS Settings
116~~~~~~~~~~~
80c0adcb 117
1ff5e4e8 118[thumbnail="screenshot/gui-create-vm-os.png"]
200114a7 119
d3c00374
TL
120When creating a virtual machine (VM), setting the proper Operating System(OS)
121allows {pve} to optimize some low level parameters. For instance Windows OS
122expect the BIOS clock to use the local time, while Unix based OS expect the
123BIOS clock to have the UTC time.
124
125[[qm_system_settings]]
126System Settings
127~~~~~~~~~~~~~~~
128
ade78a55
TL
129On VM creation you can change some basic system components of the new VM. You
130can specify which xref:qm_display[display type] you want to use.
d3c00374
TL
131[thumbnail="screenshot/gui-create-vm-system.png"]
132Additionally, the xref:qm_hard_disk[SCSI controller] can be changed.
133If you plan to install the QEMU Guest Agent, or if your selected ISO image
134already ships and installs it automatically, you may want to tick the 'Qemu
135Agent' box, which lets {pve} know that it can use its features to show some
136more information, and complete some actions (for example, shutdown or
137snapshots) more intelligently.
138
139{pve} allows to boot VMs with different firmware and machine types, namely
140xref:qm_bios_and_uefi[SeaBIOS and OVMF]. In most cases you want to switch from
3a433e9b 141the default SeaBIOS to OVMF only if you plan to use
d3c00374
TL
142xref:qm_pci_passthrough[PCIe pass through]. A VMs 'Machine Type' defines the
143hardware layout of the VM's virtual motherboard. You can choose between the
144default https://en.wikipedia.org/wiki/Intel_440FX[Intel 440FX] or the
145https://ark.intel.com/content/www/us/en/ark/products/31918/intel-82q35-graphics-and-memory-controller.html[Q35]
146chipset, which also provides a virtual PCIe bus, and thus may be desired if
5f318cc0 147one wants to pass through PCIe hardware.
5eba0743 148
80c0adcb 149[[qm_hard_disk]]
c4cba5d7
EK
150Hard Disk
151~~~~~~~~~
80c0adcb 152
3dbe1daa
TL
153[[qm_hard_disk_bus]]
154Bus/Controller
155^^^^^^^^^^^^^^
2ec49380 156Qemu can emulate a number of storage controllers:
c4cba5d7
EK
157
158* the *IDE* controller, has a design which goes back to the 1984 PC/AT disk
44f38275 159controller. Even if this controller has been superseded by recent designs,
6fb50457 160each and every OS you can think of has support for it, making it a great choice
c4cba5d7
EK
161if you want to run an OS released before 2003. You can connect up to 4 devices
162on this controller.
163
164* the *SATA* (Serial ATA) controller, dating from 2003, has a more modern
165design, allowing higher throughput and a greater number of devices to be
166connected. You can connect up to 6 devices on this controller.
167
b0b6802b
EK
168* the *SCSI* controller, designed in 1985, is commonly found on server grade
169hardware, and can connect up to 14 storage devices. {pve} emulates by default a
f4bfd701
DM
170LSI 53C895A controller.
171+
81868c7e 172A SCSI controller of type _VirtIO SCSI_ is the recommended setting if you aim for
b0b6802b
EK
173performance and is automatically selected for newly created Linux VMs since
174{pve} 4.3. Linux distributions have support for this controller since 2012, and
c4cba5d7 175FreeBSD since 2014. For Windows OSes, you need to provide an extra iso
b0b6802b
EK
176containing the drivers during the installation.
177// https://pve.proxmox.com/wiki/Paravirtualized_Block_Drivers_for_Windows#During_windows_installation.
81868c7e
EK
178If you aim at maximum performance, you can select a SCSI controller of type
179_VirtIO SCSI single_ which will allow you to select the *IO Thread* option.
180When selecting _VirtIO SCSI single_ Qemu will create a new controller for
181each disk, instead of adding all disks to the same controller.
b0b6802b 182
30e6fe00
TL
183* The *VirtIO Block* controller, often just called VirtIO or virtio-blk,
184is an older type of paravirtualized controller. It has been superseded by the
185VirtIO SCSI Controller, in terms of features.
c4cba5d7 186
1ff5e4e8 187[thumbnail="screenshot/gui-create-vm-hard-disk.png"]
3dbe1daa
TL
188
189[[qm_hard_disk_formats]]
190Image Format
191^^^^^^^^^^^^
c4cba5d7
EK
192On each controller you attach a number of emulated hard disks, which are backed
193by a file or a block device residing in the configured storage. The choice of
194a storage type will determine the format of the hard disk image. Storages which
195present block devices (LVM, ZFS, Ceph) will require the *raw disk image format*,
de14ebff 196whereas files based storages (Ext4, NFS, CIFS, GlusterFS) will let you to choose
c4cba5d7
EK
197either the *raw disk image format* or the *QEMU image format*.
198
199 * the *QEMU image format* is a copy on write format which allows snapshots, and
200 thin provisioning of the disk image.
189d3661
DC
201 * the *raw disk image* is a bit-to-bit image of a hard disk, similar to what
202 you would get when executing the `dd` command on a block device in Linux. This
4371b2fe 203 format does not support thin provisioning or snapshots by itself, requiring
30e6fe00
TL
204 cooperation from the storage layer for these tasks. It may, however, be up to
205 10% faster than the *QEMU image format*. footnote:[See this benchmark for details
43530f6f 206 https://events.static.linuxfound.org/sites/events/files/slides/CloudOpen2013_Khoa_Huynh_v3.pdf]
189d3661 207 * the *VMware image format* only makes sense if you intend to import/export the
c4cba5d7
EK
208 disk image to other hypervisors.
209
3dbe1daa
TL
210[[qm_hard_disk_cache]]
211Cache Mode
212^^^^^^^^^^
c4cba5d7
EK
213Setting the *Cache* mode of the hard drive will impact how the host system will
214notify the guest systems of block write completions. The *No cache* default
215means that the guest system will be notified that a write is complete when each
216block reaches the physical storage write queue, ignoring the host page cache.
217This provides a good balance between safety and speed.
218
219If you want the {pve} backup manager to skip a disk when doing a backup of a VM,
220you can set the *No backup* option on that disk.
221
3205ac49
EK
222If you want the {pve} storage replication mechanism to skip a disk when starting
223 a replication job, you can set the *Skip replication* option on that disk.
6fb50457 224As of {pve} 5.0, replication requires the disk images to be on a storage of type
3205ac49 225`zfspool`, so adding a disk image to other storages when the VM has replication
6fb50457 226configured requires to skip replication for this disk image.
3205ac49 227
3dbe1daa
TL
228[[qm_hard_disk_discard]]
229Trim/Discard
230^^^^^^^^^^^^
c4cba5d7 231If your storage supports _thin provisioning_ (see the storage chapter in the
53cbac40
NC
232{pve} guide), you can activate the *Discard* option on a drive. With *Discard*
233set and a _TRIM_-enabled guest OS footnote:[TRIM, UNMAP, and discard
234https://en.wikipedia.org/wiki/Trim_%28computing%29], when the VM's filesystem
235marks blocks as unused after deleting files, the controller will relay this
236information to the storage, which will then shrink the disk image accordingly.
43975153
SR
237For the guest to be able to issue _TRIM_ commands, you must enable the *Discard*
238option on the drive. Some guest operating systems may also require the
239*SSD Emulation* flag to be set. Note that *Discard* on *VirtIO Block* drives is
240only supported on guests using Linux Kernel 5.0 or higher.
c4cba5d7 241
25203dc1
NC
242If you would like a drive to be presented to the guest as a solid-state drive
243rather than a rotational hard disk, you can set the *SSD emulation* option on
244that drive. There is no requirement that the underlying storage actually be
245backed by SSDs; this feature can be used with physical media of any type.
53cbac40 246Note that *SSD emulation* is not supported on *VirtIO Block* drives.
25203dc1 247
3dbe1daa
TL
248
249[[qm_hard_disk_iothread]]
250IO Thread
251^^^^^^^^^
59552707 252The option *IO Thread* can only be used when using a disk with the
81868c7e
EK
253*VirtIO* controller, or with the *SCSI* controller, when the emulated controller
254 type is *VirtIO SCSI single*.
255With this enabled, Qemu creates one I/O thread per storage controller,
cb335798
DW
256rather than a single thread for all I/O. This can increase performance when
257multiple disks are used and each disk has its own storage controller.
c564fc52 258
80c0adcb
DM
259
260[[qm_cpu]]
34e541c5
EK
261CPU
262~~~
80c0adcb 263
1ff5e4e8 264[thumbnail="screenshot/gui-create-vm-cpu.png"]
397c74c3 265
34e541c5
EK
266A *CPU socket* is a physical slot on a PC motherboard where you can plug a CPU.
267This CPU can then contain one or many *cores*, which are independent
268processing units. Whether you have a single CPU socket with 4 cores, or two CPU
269sockets with two cores is mostly irrelevant from a performance point of view.
44f38275
TL
270However some software licenses depend on the number of sockets a machine has,
271in that case it makes sense to set the number of sockets to what the license
272allows you.
f4bfd701 273
3a433e9b 274Increasing the number of virtual CPUs (cores and sockets) will usually provide a
34e541c5 275performance improvement though that is heavily dependent on the use of the VM.
3a433e9b
OB
276Multi-threaded applications will of course benefit from a large number of
277virtual CPUs, as for each virtual cpu you add, Qemu will create a new thread of
34e541c5
EK
278execution on the host system. If you're not sure about the workload of your VM,
279it is usually a safe bet to set the number of *Total cores* to 2.
280
fb29acdd 281NOTE: It is perfectly safe if the _overall_ number of cores of all your VMs
7dd7a0b7
TL
282is greater than the number of cores on the server (e.g., 4 VMs with each 4
283cores on a machine with only 8 cores). In that case the host system will
284balance the Qemu execution threads between your server cores, just like if you
3a433e9b 285were running a standard multi-threaded application. However, {pve} will prevent
bb0fdf61
DJ
286you from starting VMs with more virtual CPU cores than physically available, as
287this will only bring the performance down due to the cost of context switches.
34e541c5 288
af54f54d
TL
289[[qm_cpu_resource_limits]]
290Resource Limits
291^^^^^^^^^^^^^^^
292
4371b2fe 293In addition to the number of virtual cores, you can configure how much resources
af54f54d
TL
294a VM can get in relation to the host CPU time and also in relation to other
295VMs.
046643ec
FG
296With the *cpulimit* (``Host CPU Time'') option you can limit how much CPU time
297the whole VM can use on the host. It is a floating point value representing CPU
af54f54d 298time in percent, so `1.0` is equal to `100%`, `2.5` to `250%` and so on. If a
4371b2fe 299single process would fully use one single core it would have `100%` CPU Time
af54f54d
TL
300usage. If a VM with four cores utilizes all its cores fully it would
301theoretically use `400%`. In reality the usage may be even a bit higher as Qemu
302can have additional threads for VM peripherals besides the vCPU core ones.
303This setting can be useful if a VM should have multiple vCPUs, as it runs a few
304processes in parallel, but the VM as a whole should not be able to run all
305vCPUs at 100% at the same time. Using a specific example: lets say we have a VM
306which would profit from having 8 vCPUs, but at no time all of those 8 cores
307should run at full load - as this would make the server so overloaded that
308other VMs and CTs would get to less CPU. So, we set the *cpulimit* limit to
309`4.0` (=400%). If all cores do the same heavy work they would all get 50% of a
310real host cores CPU time. But, if only 4 would do work they could still get
311almost 100% of a real core each.
312
313NOTE: VMs can, depending on their configuration, use additional threads e.g.,
314for networking or IO operations but also live migration. Thus a VM can show up
315to use more CPU time than just its virtual CPUs could use. To ensure that a VM
316never uses more CPU time than virtual CPUs assigned set the *cpulimit* setting
317to the same value as the total core count.
318
319The second CPU resource limiting setting, *cpuunits* (nowadays often called CPU
320shares or CPU weight), controls how much CPU time a VM gets in regards to other
321VMs running. It is a relative weight which defaults to `1024`, if you increase
322this for a VM it will be prioritized by the scheduler in comparison to other
323VMs with lower weight. E.g., if VM 100 has set the default 1024 and VM 200 was
324changed to `2048`, the latter VM 200 would receive twice the CPU bandwidth than
325the first VM 100.
326
327For more information see `man systemd.resource-control`, here `CPUQuota`
328corresponds to `cpulimit` and `CPUShares` corresponds to our `cpuunits`
329setting, visit its Notes section for references and implementation details.
330
331CPU Type
332^^^^^^^^
333
34e541c5
EK
334Qemu can emulate a number different of *CPU types* from 486 to the latest Xeon
335processors. Each new processor generation adds new features, like hardware
336assisted 3d rendering, random number generation, memory protection, etc ...
337Usually you should select for your VM a processor type which closely matches the
338CPU of the host system, as it means that the host CPU features (also called _CPU
339flags_ ) will be available in your VMs. If you want an exact match, you can set
340the CPU type to *host* in which case the VM will have exactly the same CPU flags
f4bfd701
DM
341as your host system.
342
34e541c5
EK
343This has a downside though. If you want to do a live migration of VMs between
344different hosts, your VM might end up on a new system with a different CPU type.
345If the CPU flags passed to the guest are missing, the qemu process will stop. To
346remedy this Qemu has also its own CPU type *kvm64*, that {pve} uses by defaults.
347kvm64 is a Pentium 4 look a like CPU type, which has a reduced CPU flags set,
f4bfd701
DM
348but is guaranteed to work everywhere.
349
350In short, if you care about live migration and moving VMs between nodes, leave
af54f54d
TL
351the kvm64 default. If you don’t care about live migration or have a homogeneous
352cluster where all nodes have the same CPU, set the CPU type to host, as in
353theory this will give your guests maximum performance.
354
9e797d8c
SR
355Custom CPU Types
356^^^^^^^^^^^^^^^^
357
358You can specify custom CPU types with a configurable set of features. These are
359maintained in the configuration file `/etc/pve/virtual-guest/cpu-models.conf` by
360an administrator. See `man cpu-models.conf` for format details.
361
362Specified custom types can be selected by any user with the `Sys.Audit`
363privilege on `/nodes`. When configuring a custom CPU type for a VM via the CLI
364or API, the name needs to be prefixed with 'custom-'.
365
72ae8aa2
FG
366Meltdown / Spectre related CPU flags
367^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^
368
2975cb7a 369There are several CPU flags related to the Meltdown and Spectre vulnerabilities
72ae8aa2
FG
370footnote:[Meltdown Attack https://meltdownattack.com/] which need to be set
371manually unless the selected CPU type of your VM already enables them by default.
372
2975cb7a 373There are two requirements that need to be fulfilled in order to use these
72ae8aa2 374CPU flags:
5dba2677 375
72ae8aa2
FG
376* The host CPU(s) must support the feature and propagate it to the guest's virtual CPU(s)
377* The guest operating system must be updated to a version which mitigates the
378 attacks and is able to utilize the CPU feature
379
2975cb7a
AD
380Otherwise you need to set the desired CPU flag of the virtual CPU, either by
381editing the CPU options in the WebUI, or by setting the 'flags' property of the
382'cpu' option in the VM configuration file.
383
384For Spectre v1,v2,v4 fixes, your CPU or system vendor also needs to provide a
72ae8aa2
FG
385so-called ``microcode update'' footnote:[You can use `intel-microcode' /
386`amd-microcode' from Debian non-free if your vendor does not provide such an
387update. Note that not all affected CPUs can be updated to support spec-ctrl.]
388for your CPU.
5dba2677 389
2975cb7a
AD
390
391To check if the {pve} host is vulnerable, execute the following command as root:
5dba2677
TL
392
393----
2975cb7a 394for f in /sys/devices/system/cpu/vulnerabilities/*; do echo "${f##*/} -" $(cat "$f"); done
5dba2677
TL
395----
396
144d5ede 397A community script is also available to detect is the host is still vulnerable.
2975cb7a 398footnote:[spectre-meltdown-checker https://meltdown.ovh/]
72ae8aa2 399
2975cb7a
AD
400Intel processors
401^^^^^^^^^^^^^^^^
72ae8aa2 402
2975cb7a
AD
403* 'pcid'
404+
144d5ede 405This reduces the performance impact of the Meltdown (CVE-2017-5754) mitigation
2975cb7a
AD
406called 'Kernel Page-Table Isolation (KPTI)', which effectively hides
407the Kernel memory from the user space. Without PCID, KPTI is quite an expensive
408mechanism footnote:[PCID is now a critical performance/security feature on x86
409https://groups.google.com/forum/m/#!topic/mechanical-sympathy/L9mHTbeQLNU].
410+
411To check if the {pve} host supports PCID, execute the following command as root:
412+
72ae8aa2 413----
2975cb7a 414# grep ' pcid ' /proc/cpuinfo
72ae8aa2 415----
2975cb7a
AD
416+
417If this does not return empty your host's CPU has support for 'pcid'.
72ae8aa2 418
2975cb7a
AD
419* 'spec-ctrl'
420+
144d5ede
WB
421Required to enable the Spectre v1 (CVE-2017-5753) and Spectre v2 (CVE-2017-5715) fix,
422in cases where retpolines are not sufficient.
423Included by default in Intel CPU models with -IBRS suffix.
424Must be explicitly turned on for Intel CPU models without -IBRS suffix.
425Requires an updated host CPU microcode (intel-microcode >= 20180425).
2975cb7a
AD
426+
427* 'ssbd'
428+
144d5ede
WB
429Required to enable the Spectre V4 (CVE-2018-3639) fix. Not included by default in any Intel CPU model.
430Must be explicitly turned on for all Intel CPU models.
431Requires an updated host CPU microcode(intel-microcode >= 20180703).
72ae8aa2 432
72ae8aa2 433
2975cb7a
AD
434AMD processors
435^^^^^^^^^^^^^^
436
437* 'ibpb'
438+
144d5ede
WB
439Required to enable the Spectre v1 (CVE-2017-5753) and Spectre v2 (CVE-2017-5715) fix,
440in cases where retpolines are not sufficient.
441Included by default in AMD CPU models with -IBPB suffix.
442Must be explicitly turned on for AMD CPU models without -IBPB suffix.
2975cb7a
AD
443Requires the host CPU microcode to support this feature before it can be used for guest CPUs.
444
445
446
447* 'virt-ssbd'
448+
449Required to enable the Spectre v4 (CVE-2018-3639) fix.
144d5ede
WB
450Not included by default in any AMD CPU model.
451Must be explicitly turned on for all AMD CPU models.
452This should be provided to guests, even if amd-ssbd is also provided, for maximum guest compatibility.
453Note that this must be explicitly enabled when when using the "host" cpu model,
454because this is a virtual feature which does not exist in the physical CPUs.
2975cb7a
AD
455
456
457* 'amd-ssbd'
458+
144d5ede
WB
459Required to enable the Spectre v4 (CVE-2018-3639) fix.
460Not included by default in any AMD CPU model. Must be explicitly turned on for all AMD CPU models.
461This provides higher performance than virt-ssbd, therefore a host supporting this should always expose this to guests if possible.
2975cb7a
AD
462virt-ssbd should none the less also be exposed for maximum guest compatibility as some kernels only know about virt-ssbd.
463
464
465* 'amd-no-ssb'
466+
467Recommended to indicate the host is not vulnerable to Spectre V4 (CVE-2018-3639).
144d5ede
WB
468Not included by default in any AMD CPU model.
469Future hardware generations of CPU will not be vulnerable to CVE-2018-3639,
470and thus the guest should be told not to enable its mitigations, by exposing amd-no-ssb.
2975cb7a
AD
471This is mutually exclusive with virt-ssbd and amd-ssbd.
472
5dba2677 473
af54f54d
TL
474NUMA
475^^^^
476You can also optionally emulate a *NUMA*
477footnote:[https://en.wikipedia.org/wiki/Non-uniform_memory_access] architecture
478in your VMs. The basics of the NUMA architecture mean that instead of having a
479global memory pool available to all your cores, the memory is spread into local
480banks close to each socket.
34e541c5
EK
481This can bring speed improvements as the memory bus is not a bottleneck
482anymore. If your system has a NUMA architecture footnote:[if the command
483`numactl --hardware | grep available` returns more than one node, then your host
484system has a NUMA architecture] we recommend to activate the option, as this
af54f54d
TL
485will allow proper distribution of the VM resources on the host system.
486This option is also required to hot-plug cores or RAM in a VM.
34e541c5
EK
487
488If the NUMA option is used, it is recommended to set the number of sockets to
4ccb911c 489the number of nodes of the host system.
34e541c5 490
af54f54d
TL
491vCPU hot-plug
492^^^^^^^^^^^^^
493
494Modern operating systems introduced the capability to hot-plug and, to a
3a433e9b 495certain extent, hot-unplug CPUs in a running system. Virtualization allows us
4371b2fe
FG
496to avoid a lot of the (physical) problems real hardware can cause in such
497scenarios.
498Still, this is a rather new and complicated feature, so its use should be
499restricted to cases where its absolutely needed. Most of the functionality can
500be replicated with other, well tested and less complicated, features, see
af54f54d
TL
501xref:qm_cpu_resource_limits[Resource Limits].
502
503In {pve} the maximal number of plugged CPUs is always `cores * sockets`.
504To start a VM with less than this total core count of CPUs you may use the
4371b2fe 505*vpus* setting, it denotes how many vCPUs should be plugged in at VM start.
af54f54d 506
4371b2fe 507Currently only this feature is only supported on Linux, a kernel newer than 3.10
af54f54d
TL
508is needed, a kernel newer than 4.7 is recommended.
509
510You can use a udev rule as follow to automatically set new CPUs as online in
511the guest:
512
513----
514SUBSYSTEM=="cpu", ACTION=="add", TEST=="online", ATTR{online}=="0", ATTR{online}="1"
515----
516
517Save this under /etc/udev/rules.d/ as a file ending in `.rules`.
518
519Note: CPU hot-remove is machine dependent and requires guest cooperation.
520The deletion command does not guarantee CPU removal to actually happen,
521typically it's a request forwarded to guest using target dependent mechanism,
522e.g., ACPI on x86/amd64.
523
80c0adcb
DM
524
525[[qm_memory]]
34e541c5
EK
526Memory
527~~~~~~
80c0adcb 528
34e541c5
EK
529For each VM you have the option to set a fixed size memory or asking
530{pve} to dynamically allocate memory based on the current RAM usage of the
59552707 531host.
34e541c5 532
96124d0f 533.Fixed Memory Allocation
1ff5e4e8 534[thumbnail="screenshot/gui-create-vm-memory.png"]
96124d0f 535
9ea21953 536When setting memory and minimum memory to the same amount
9fb002e6 537{pve} will simply allocate what you specify to your VM.
34e541c5 538
9abfec65
DC
539Even when using a fixed memory size, the ballooning device gets added to the
540VM, because it delivers useful information such as how much memory the guest
541really uses.
542In general, you should leave *ballooning* enabled, but if you want to disable
e60ce90c 543it (e.g. for debugging purposes), simply uncheck
9fb002e6 544*Ballooning Device* or set
9abfec65
DC
545
546 balloon: 0
547
548in the configuration.
549
96124d0f 550.Automatic Memory Allocation
96124d0f 551
34e541c5 552// see autoballoon() in pvestatd.pm
58e04593 553When setting the minimum memory lower than memory, {pve} will make sure that the
34e541c5
EK
554minimum amount you specified is always available to the VM, and if RAM usage on
555the host is below 80%, will dynamically add memory to the guest up to the
f4bfd701
DM
556maximum memory specified.
557
a35aad4a 558When the host is running low on RAM, the VM will then release some memory
34e541c5
EK
559back to the host, swapping running processes if needed and starting the oom
560killer in last resort. The passing around of memory between host and guest is
561done via a special `balloon` kernel driver running inside the guest, which will
562grab or release memory pages from the host.
563footnote:[A good explanation of the inner workings of the balloon driver can be found here https://rwmj.wordpress.com/2010/07/17/virtio-balloon/]
564
c9f6e1a4
EK
565When multiple VMs use the autoallocate facility, it is possible to set a
566*Shares* coefficient which indicates the relative amount of the free host memory
470d4313 567that each VM should take. Suppose for instance you have four VMs, three of them
a35aad4a 568running an HTTP server and the last one is a database server. To cache more
c9f6e1a4
EK
569database blocks in the database server RAM, you would like to prioritize the
570database VM when spare RAM is available. For this you assign a Shares property
571of 3000 to the database VM, leaving the other VMs to the Shares default setting
470d4313 572of 1000. The host server has 32GB of RAM, and is currently using 16GB, leaving 32
c9f6e1a4
EK
573* 80/100 - 16 = 9GB RAM to be allocated to the VMs. The database VM will get 9 *
5743000 / (3000 + 1000 + 1000 + 1000) = 4.5 GB extra RAM and each HTTP server will
a35aad4a 575get 1.5 GB.
c9f6e1a4 576
34e541c5
EK
577All Linux distributions released after 2010 have the balloon kernel driver
578included. For Windows OSes, the balloon driver needs to be added manually and can
579incur a slowdown of the guest, so we don't recommend using it on critical
59552707 580systems.
34e541c5
EK
581// see https://forum.proxmox.com/threads/solved-hyper-threading-vs-no-hyper-threading-fixed-vs-variable-memory.20265/
582
470d4313 583When allocating RAM to your VMs, a good rule of thumb is always to leave 1GB
34e541c5
EK
584of RAM available to the host.
585
80c0adcb
DM
586
587[[qm_network_device]]
1ff7835b
EK
588Network Device
589~~~~~~~~~~~~~~
80c0adcb 590
1ff5e4e8 591[thumbnail="screenshot/gui-create-vm-network.png"]
c24ddb0a 592
1ff7835b
EK
593Each VM can have many _Network interface controllers_ (NIC), of four different
594types:
595
596 * *Intel E1000* is the default, and emulates an Intel Gigabit network card.
597 * the *VirtIO* paravirtualized NIC should be used if you aim for maximum
598performance. Like all VirtIO devices, the guest OS should have the proper driver
599installed.
600 * the *Realtek 8139* emulates an older 100 MB/s network card, and should
59552707 601only be used when emulating older operating systems ( released before 2002 )
1ff7835b
EK
602 * the *vmxnet3* is another paravirtualized device, which should only be used
603when importing a VM from another hypervisor.
604
605{pve} will generate for each NIC a random *MAC address*, so that your VM is
606addressable on Ethernet networks.
607
470d4313 608The NIC you added to the VM can follow one of two different models:
af9c6de1
EK
609
610 * in the default *Bridged mode* each virtual NIC is backed on the host by a
611_tap device_, ( a software loopback device simulating an Ethernet NIC ). This
612tap device is added to a bridge, by default vmbr0 in {pve}. In this mode, VMs
613have direct access to the Ethernet LAN on which the host is located.
614 * in the alternative *NAT mode*, each virtual NIC will only communicate with
470d4313
DC
615the Qemu user networking stack, where a built-in router and DHCP server can
616provide network access. This built-in DHCP will serve addresses in the private
af9c6de1 61710.0.2.0/24 range. The NAT mode is much slower than the bridged mode, and
f5041150
DC
618should only be used for testing. This mode is only available via CLI or the API,
619but not via the WebUI.
af9c6de1
EK
620
621You can also skip adding a network device when creating a VM by selecting *No
622network device*.
623
624.Multiqueue
1ff7835b 625If you are using the VirtIO driver, you can optionally activate the
af9c6de1 626*Multiqueue* option. This option allows the guest OS to process networking
1ff7835b 627packets using multiple virtual CPUs, providing an increase in the total number
470d4313 628of packets transferred.
1ff7835b
EK
629
630//http://blog.vmsplice.net/2011/09/qemu-internals-vhost-architecture.html
631When using the VirtIO driver with {pve}, each NIC network queue is passed to the
a35aad4a 632host kernel, where the queue will be processed by a kernel thread spawned by the
1ff7835b
EK
633vhost driver. With this option activated, it is possible to pass _multiple_
634network queues to the host kernel for each NIC.
635
636//https://access.redhat.com/documentation/en-US/Red_Hat_Enterprise_Linux/7/html/Virtualization_Tuning_and_Optimization_Guide/sect-Virtualization_Tuning_Optimization_Guide-Networking-Techniques.html#sect-Virtualization_Tuning_Optimization_Guide-Networking-Multi-queue_virtio-net
af9c6de1 637When using Multiqueue, it is recommended to set it to a value equal
1ff7835b
EK
638to the number of Total Cores of your guest. You also need to set in
639the VM the number of multi-purpose channels on each VirtIO NIC with the ethtool
59552707 640command:
1ff7835b 641
7a0d4784 642`ethtool -L ens1 combined X`
1ff7835b
EK
643
644where X is the number of the number of vcpus of the VM.
645
af9c6de1 646You should note that setting the Multiqueue parameter to a value greater
1ff7835b
EK
647than one will increase the CPU load on the host and guest systems as the
648traffic increases. We recommend to set this option only when the VM has to
649process a great number of incoming connections, such as when the VM is running
650as a router, reverse proxy or a busy HTTP server doing long polling.
651
6cb67d7f
DC
652[[qm_display]]
653Display
654~~~~~~~
655
656QEMU can virtualize a few types of VGA hardware. Some examples are:
657
658* *std*, the default, emulates a card with Bochs VBE extensions.
1368dc02
TL
659* *cirrus*, this was once the default, it emulates a very old hardware module
660with all its problems. This display type should only be used if really
661necessary footnote:[https://www.kraxel.org/blog/2014/10/qemu-using-cirrus-considered-harmful/
662qemu: using cirrus considered harmful], e.g., if using Windows XP or earlier
6cb67d7f
DC
663* *vmware*, is a VMWare SVGA-II compatible adapter.
664* *qxl*, is the QXL paravirtualized graphics card. Selecting this also
37422176
AL
665enables https://www.spice-space.org/[SPICE] (a remote viewer protocol) for the
666VM.
6cb67d7f
DC
667
668You can edit the amount of memory given to the virtual GPU, by setting
1368dc02 669the 'memory' option. This can enable higher resolutions inside the VM,
6cb67d7f
DC
670especially with SPICE/QXL.
671
1368dc02
TL
672As the memory is reserved by display device, selecting Multi-Monitor mode
673for SPICE (e.g., `qxl2` for dual monitors) has some implications:
6cb67d7f 674
1368dc02
TL
675* Windows needs a device for each monitor, so if your 'ostype' is some
676version of Windows, {pve} gives the VM an extra device per monitor.
6cb67d7f 677Each device gets the specified amount of memory.
1368dc02 678
6cb67d7f
DC
679* Linux VMs, can always enable more virtual monitors, but selecting
680a Multi-Monitor mode multiplies the memory given to the device with
681the number of monitors.
682
1368dc02
TL
683Selecting `serialX` as display 'type' disables the VGA output, and redirects
684the Web Console to the selected serial port. A configured display 'memory'
685setting will be ignored in that case.
80c0adcb 686
dbb44ef0 687[[qm_usb_passthrough]]
685cc8e0
DC
688USB Passthrough
689~~~~~~~~~~~~~~~
80c0adcb 690
685cc8e0
DC
691There are two different types of USB passthrough devices:
692
470d4313 693* Host USB passthrough
685cc8e0
DC
694* SPICE USB passthrough
695
696Host USB passthrough works by giving a VM a USB device of the host.
697This can either be done via the vendor- and product-id, or
698via the host bus and port.
699
700The vendor/product-id looks like this: *0123:abcd*,
701where *0123* is the id of the vendor, and *abcd* is the id
702of the product, meaning two pieces of the same usb device
703have the same id.
704
705The bus/port looks like this: *1-2.3.4*, where *1* is the bus
706and *2.3.4* is the port path. This represents the physical
707ports of your host (depending of the internal order of the
708usb controllers).
709
710If a device is present in a VM configuration when the VM starts up,
711but the device is not present in the host, the VM can boot without problems.
470d4313 712As soon as the device/port is available in the host, it gets passed through.
685cc8e0 713
e60ce90c 714WARNING: Using this kind of USB passthrough means that you cannot move
685cc8e0
DC
715a VM online to another host, since the hardware is only available
716on the host the VM is currently residing.
717
718The second type of passthrough is SPICE USB passthrough. This is useful
719if you use a SPICE client which supports it. If you add a SPICE USB port
720to your VM, you can passthrough a USB device from where your SPICE client is,
721directly to the VM (for example an input device or hardware dongle).
722
80c0adcb
DM
723
724[[qm_bios_and_uefi]]
076d60ae
DC
725BIOS and UEFI
726~~~~~~~~~~~~~
727
728In order to properly emulate a computer, QEMU needs to use a firmware.
55ce3375
TL
729Which, on common PCs often known as BIOS or (U)EFI, is executed as one of the
730first steps when booting a VM. It is responsible for doing basic hardware
731initialization and for providing an interface to the firmware and hardware for
732the operating system. By default QEMU uses *SeaBIOS* for this, which is an
733open-source, x86 BIOS implementation. SeaBIOS is a good choice for most
734standard setups.
076d60ae 735
8e5720fd
SR
736Some operating systems (such as Windows 11) may require use of an UEFI
737compatible implementation instead. In such cases, you must rather use *OVMF*,
738which is an open-source UEFI implementation. footnote:[See the OVMF Project https://github.com/tianocore/tianocore.github.io/wiki/OVMF]
739
740There are other scenarios in which a BIOS is not a good firmware to boot from,
741e.g. if you want to do VGA passthrough. footnote:[Alex Williamson has a very
742good blog entry about this https://vfio.blogspot.co.at/2014/08/primary-graphics-assignment-without-vga.html]
076d60ae
DC
743
744If you want to use OVMF, there are several things to consider:
745
746In order to save things like the *boot order*, there needs to be an EFI Disk.
747This disk will be included in backups and snapshots, and there can only be one.
748
749You can create such a disk with the following command:
750
8e5720fd 751 qm set <vmid> -efidisk0 <storage>:1,format=<format>,efitype=4m,pre-enrolled-keys=1
076d60ae
DC
752
753Where *<storage>* is the storage where you want to have the disk, and
754*<format>* is a format which the storage supports. Alternatively, you can
755create such a disk through the web interface with 'Add' -> 'EFI Disk' in the
756hardware section of a VM.
757
8e5720fd
SR
758The *efitype* option specifies which version of the OVMF firmware should be
759used. For new VMs, this should always be '4m', as it supports Secure Boot and
760has more space allocated to support future development (this is the default in
761the GUI).
762
763*pre-enroll-keys* specifies if the efidisk should come pre-loaded with
764distribution-specific and Microsoft Standard Secure Boot keys. It also enables
765Secure Boot by default (though it can still be disabled in the OVMF menu within
766the VM).
767
768NOTE: If you want to start using Secure Boot in an existing VM (that still uses
769a '2m' efidisk), you need to recreate the efidisk. To do so, delete the old one
770(`qm set <vmid> -delete efidisk0`) and add a new one as described above. This
771will reset any custom configurations you have made in the OVMF menu!
772
076d60ae 773When using OVMF with a virtual display (without VGA passthrough),
8e5720fd 774you need to set the client resolution in the OVMF menu (which you can reach
076d60ae
DC
775with a press of the ESC button during boot), or you have to choose
776SPICE as the display type.
777
95e8e1b7
SR
778[[qm_tpm]]
779Trusted Platform Module (TPM)
780~~~~~~~~~~~~~~~~~~~~~~~~~~~~~
781
782A *Trusted Platform Module* is a device which stores secret data - such as
783encryption keys - securely and provides tamper-resistance functions for
784validating system boot.
785
786Certain operating systems (e.g. Windows 11) require such a device to be attached
787to a machine (be it physical or virtual).
788
789A TPM is added by specifying a *tpmstate* volume. This works similar to an
790efidisk, in that it cannot be changed (only removed) once created. You can add
791one via the following command:
792
793 qm set <vmid> -tpmstate0 <storage>:1,version=<version>
794
795Where *<storage>* is the storage you want to put the state on, and *<version>*
796is either 'v1.2' or 'v2.0'. You can also add one via the web interface, by
797choosing 'Add' -> 'TPM State' in the hardware section of a VM.
798
799The 'v2.0' TPM spec is newer and better supported, so unless you have a specific
800implementation that requires a 'v1.2' TPM, it should be preferred.
801
802NOTE: Compared to a physical TPM, an emulated one does *not* provide any real
803security benefits. The point of a TPM is that the data on it cannot be modified
804easily, except via commands specified as part of the TPM spec. Since with an
805emulated device the data storage happens on a regular volume, it can potentially
806be edited by anyone with access to it.
807
0ad30983
DC
808[[qm_ivshmem]]
809Inter-VM shared memory
810~~~~~~~~~~~~~~~~~~~~~~
811
8861c7ad
TL
812You can add an Inter-VM shared memory device (`ivshmem`), which allows one to
813share memory between the host and a guest, or also between multiple guests.
0ad30983
DC
814
815To add such a device, you can use `qm`:
816
817 qm set <vmid> -ivshmem size=32,name=foo
818
819Where the size is in MiB. The file will be located under
820`/dev/shm/pve-shm-$name` (the default name is the vmid).
821
4d1a19eb
TL
822NOTE: Currently the device will get deleted as soon as any VM using it got
823shutdown or stopped. Open connections will still persist, but new connections
824to the exact same device cannot be made anymore.
825
8861c7ad 826A use case for such a device is the Looking Glass
451bb75f
SR
827footnote:[Looking Glass: https://looking-glass.io/] project, which enables high
828performance, low-latency display mirroring between host and guest.
0ad30983 829
ca8c3009
AL
830[[qm_audio_device]]
831Audio Device
832~~~~~~~~~~~~
833
834To add an audio device run the following command:
835
836----
837qm set <vmid> -audio0 device=<device>
838----
839
840Supported audio devices are:
841
842* `ich9-intel-hda`: Intel HD Audio Controller, emulates ICH9
843* `intel-hda`: Intel HD Audio Controller, emulates ICH6
844* `AC97`: Audio Codec '97, useful for older operating systems like Windows XP
845
cf41761d
AL
846There are two backends available:
847
848* 'spice'
849* 'none'
850
851The 'spice' backend can be used in combination with xref:qm_display[SPICE] while
852the 'none' backend can be useful if an audio device is needed in the VM for some
853software to work. To use the physical audio device of the host use device
854passthrough (see xref:qm_pci_passthrough[PCI Passthrough] and
855xref:qm_usb_passthrough[USB Passthrough]). Remote protocols like Microsoft’s RDP
856have options to play sound.
857
ca8c3009 858
adb2c91d
SR
859[[qm_virtio_rng]]
860VirtIO RNG
861~~~~~~~~~~
862
863A RNG (Random Number Generator) is a device providing entropy ('randomness') to
864a system. A virtual hardware-RNG can be used to provide such entropy from the
865host system to a guest VM. This helps to avoid entropy starvation problems in
866the guest (a situation where not enough entropy is available and the system may
867slow down or run into problems), especially during the guests boot process.
868
869To add a VirtIO-based emulated RNG, run the following command:
870
871----
872qm set <vmid> -rng0 source=<source>[,max_bytes=X,period=Y]
873----
874
875`source` specifies where entropy is read from on the host and has to be one of
876the following:
877
878* `/dev/urandom`: Non-blocking kernel entropy pool (preferred)
879* `/dev/random`: Blocking kernel pool (not recommended, can lead to entropy
880 starvation on the host system)
881* `/dev/hwrng`: To pass through a hardware RNG attached to the host (if multiple
882 are available, the one selected in
883 `/sys/devices/virtual/misc/hw_random/rng_current` will be used)
884
885A limit can be specified via the `max_bytes` and `period` parameters, they are
886read as `max_bytes` per `period` in milliseconds. However, it does not represent
887a linear relationship: 1024B/1000ms would mean that up to 1 KiB of data becomes
888available on a 1 second timer, not that 1 KiB is streamed to the guest over the
889course of one second. Reducing the `period` can thus be used to inject entropy
890into the guest at a faster rate.
891
892By default, the limit is set to 1024 bytes per 1000 ms (1 KiB/s). It is
893recommended to always use a limiter to avoid guests using too many host
894resources. If desired, a value of '0' for `max_bytes` can be used to disable
895all limits.
896
777cf894 897[[qm_bootorder]]
8cd6f474
TL
898Device Boot Order
899~~~~~~~~~~~~~~~~~
777cf894
SR
900
901QEMU can tell the guest which devices it should boot from, and in which order.
902This can be specified in the config via the `boot` property, e.g.:
903
904----
905boot: order=scsi0;net0;hostpci0
906----
907
908[thumbnail="screenshot/gui-qemu-edit-bootorder.png"]
909
910This way, the guest would first attempt to boot from the disk `scsi0`, if that
911fails, it would go on to attempt network boot from `net0`, and in case that
912fails too, finally attempt to boot from a passed through PCIe device (seen as
913disk in case of NVMe, otherwise tries to launch into an option ROM).
914
915On the GUI you can use a drag-and-drop editor to specify the boot order, and use
916the checkbox to enable or disable certain devices for booting altogether.
917
918NOTE: If your guest uses multiple disks to boot the OS or load the bootloader,
919all of them must be marked as 'bootable' (that is, they must have the checkbox
920enabled or appear in the list in the config) for the guest to be able to boot.
921This is because recent SeaBIOS and OVMF versions only initialize disks if they
922are marked 'bootable'.
923
924In any case, even devices not appearing in the list or having the checkmark
925disabled will still be available to the guest, once it's operating system has
926booted and initialized them. The 'bootable' flag only affects the guest BIOS and
927bootloader.
928
929
288e3f46
EK
930[[qm_startup_and_shutdown]]
931Automatic Start and Shutdown of Virtual Machines
932~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~
933
934After creating your VMs, you probably want them to start automatically
935when the host system boots. For this you need to select the option 'Start at
936boot' from the 'Options' Tab of your VM in the web interface, or set it with
937the following command:
938
939 qm set <vmid> -onboot 1
940
4dbeb548
DM
941.Start and Shutdown Order
942
1ff5e4e8 943[thumbnail="screenshot/gui-qemu-edit-start-order.png"]
4dbeb548
DM
944
945In some case you want to be able to fine tune the boot order of your
946VMs, for instance if one of your VM is providing firewalling or DHCP
947to other guest systems. For this you can use the following
948parameters:
288e3f46
EK
949
950* *Start/Shutdown order*: Defines the start order priority. E.g. set it to 1 if
951you want the VM to be the first to be started. (We use the reverse startup
952order for shutdown, so a machine with a start order of 1 would be the last to
7eed72d8 953be shut down). If multiple VMs have the same order defined on a host, they will
d750c851 954additionally be ordered by 'VMID' in ascending order.
288e3f46
EK
955* *Startup delay*: Defines the interval between this VM start and subsequent
956VMs starts . E.g. set it to 240 if you want to wait 240 seconds before starting
957other VMs.
958* *Shutdown timeout*: Defines the duration in seconds {pve} should wait
959for the VM to be offline after issuing a shutdown command.
7eed72d8 960By default this value is set to 180, which means that {pve} will issue a
d750c851
WB
961shutdown request and wait 180 seconds for the machine to be offline. If
962the machine is still online after the timeout it will be stopped forcefully.
288e3f46 963
2b2c6286
TL
964NOTE: VMs managed by the HA stack do not follow the 'start on boot' and
965'boot order' options currently. Those VMs will be skipped by the startup and
966shutdown algorithm as the HA manager itself ensures that VMs get started and
967stopped.
968
288e3f46 969Please note that machines without a Start/Shutdown order parameter will always
7eed72d8 970start after those where the parameter is set. Further, this parameter can only
d750c851 971be enforced between virtual machines running on the same host, not
288e3f46 972cluster-wide.
076d60ae 973
0f7778ac
DW
974If you require a delay between the host boot and the booting of the first VM,
975see the section on xref:first_guest_boot_delay[Proxmox VE Node Management].
976
c0f039aa
AL
977
978[[qm_qemu_agent]]
979Qemu Guest Agent
980~~~~~~~~~~~~~~~~
981
982The Qemu Guest Agent is a service which runs inside the VM, providing a
983communication channel between the host and the guest. It is used to exchange
984information and allows the host to issue commands to the guest.
985
986For example, the IP addresses in the VM summary panel are fetched via the guest
987agent.
988
989Or when starting a backup, the guest is told via the guest agent to sync
990outstanding writes via the 'fs-freeze' and 'fs-thaw' commands.
991
992For the guest agent to work properly the following steps must be taken:
993
994* install the agent in the guest and make sure it is running
995* enable the communication via the agent in {pve}
996
997Install Guest Agent
998^^^^^^^^^^^^^^^^^^^
999
1000For most Linux distributions, the guest agent is available. The package is
1001usually named `qemu-guest-agent`.
1002
1003For Windows, it can be installed from the
1004https://fedorapeople.org/groups/virt/virtio-win/direct-downloads/stable-virtio/virtio-win.iso[Fedora
1005VirtIO driver ISO].
1006
1007Enable Guest Agent Communication
1008^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^
1009
1010Communication from {pve} with the guest agent can be enabled in the VM's
1011*Options* panel. A fresh start of the VM is necessary for the changes to take
1012effect.
1013
1014It is possible to enable the 'Run guest-trim' option. With this enabled,
1015{pve} will issue a trim command to the guest after the following
1016operations that have the potential to write out zeros to the storage:
1017
1018* moving a disk to another storage
1019* live migrating a VM to another node with local storage
1020
1021On a thin provisioned storage, this can help to free up unused space.
1022
1023Troubleshooting
1024^^^^^^^^^^^^^^^
1025
1026.VM does not shut down
1027
1028Make sure the guest agent is installed and running.
1029
1030Once the guest agent is enabled, {pve} will send power commands like
1031'shutdown' via the guest agent. If the guest agent is not running, commands
1032cannot get executed properly and the shutdown command will run into a timeout.
1033
22a0091c
AL
1034[[qm_spice_enhancements]]
1035SPICE Enhancements
1036~~~~~~~~~~~~~~~~~~
1037
1038SPICE Enhancements are optional features that can improve the remote viewer
1039experience.
1040
1041To enable them via the GUI go to the *Options* panel of the virtual machine. Run
1042the following command to enable them via the CLI:
1043
1044----
1045qm set <vmid> -spice_enhancements foldersharing=1,videostreaming=all
1046----
1047
1048NOTE: To use these features the <<qm_display,*Display*>> of the virtual machine
1049must be set to SPICE (qxl).
1050
1051Folder Sharing
1052^^^^^^^^^^^^^^
1053
1054Share a local folder with the guest. The `spice-webdavd` daemon needs to be
1055installed in the guest. It makes the shared folder available through a local
1056WebDAV server located at http://localhost:9843.
1057
1058For Windows guests the installer for the 'Spice WebDAV daemon' can be downloaded
1059from the
1060https://www.spice-space.org/download.html#windows-binaries[official SPICE website].
1061
1062Most Linux distributions have a package called `spice-webdavd` that can be
1063installed.
1064
1065To share a folder in Virt-Viewer (Remote Viewer) go to 'File -> Preferences'.
1066Select the folder to share and then enable the checkbox.
1067
1068NOTE: Folder sharing currently only works in the Linux version of Virt-Viewer.
1069
0dcd22f5
AL
1070CAUTION: Experimental! Currently this feature does not work reliably.
1071
22a0091c
AL
1072Video Streaming
1073^^^^^^^^^^^^^^^
1074
1075Fast refreshing areas are encoded into a video stream. Two options exist:
1076
1077* *all*: Any fast refreshing area will be encoded into a video stream.
1078* *filter*: Additional filters are used to decide if video streaming should be
1079 used (currently only small window surfaces are skipped).
1080
1081A general recommendation if video streaming should be enabled and which option
1082to choose from cannot be given. Your mileage may vary depending on the specific
1083circumstances.
1084
1085Troubleshooting
1086^^^^^^^^^^^^^^^
1087
19a58e02 1088.Shared folder does not show up
22a0091c
AL
1089
1090Make sure the WebDAV service is enabled and running in the guest. On Windows it
1091is called 'Spice webdav proxy'. In Linux the name is 'spice-webdavd' but can be
1092different depending on the distribution.
1093
1094If the service is running, check the WebDAV server by opening
1095http://localhost:9843 in a browser in the guest.
1096
1097It can help to restart the SPICE session.
c73c190f
DM
1098
1099[[qm_migration]]
1100Migration
1101---------
1102
1ff5e4e8 1103[thumbnail="screenshot/gui-qemu-migrate.png"]
e4bcef0a 1104
c73c190f
DM
1105If you have a cluster, you can migrate your VM to another host with
1106
1107 qm migrate <vmid> <target>
1108
8df8cfb7
DC
1109There are generally two mechanisms for this
1110
1111* Online Migration (aka Live Migration)
1112* Offline Migration
1113
1114Online Migration
1115~~~~~~~~~~~~~~~~
1116
c73c190f
DM
1117When your VM is running and it has no local resources defined (such as disks
1118on local storage, passed through devices, etc.) you can initiate a live
1119migration with the -online flag.
1120
8df8cfb7
DC
1121How it works
1122^^^^^^^^^^^^
1123
1124This starts a Qemu Process on the target host with the 'incoming' flag, which
1125means that the process starts and waits for the memory data and device states
1126from the source Virtual Machine (since all other resources, e.g. disks,
1127are shared, the memory content and device state are the only things left
1128to transmit).
1129
1130Once this connection is established, the source begins to send the memory
1131content asynchronously to the target. If the memory on the source changes,
1132those sections are marked dirty and there will be another pass of sending data.
1133This happens until the amount of data to send is so small that it can
1134pause the VM on the source, send the remaining data to the target and start
1135the VM on the target in under a second.
1136
1137Requirements
1138^^^^^^^^^^^^
1139
1140For Live Migration to work, there are some things required:
1141
1142* The VM has no local resources (e.g. passed through devices, local disks, etc.)
1143* The hosts are in the same {pve} cluster.
1144* The hosts have a working (and reliable) network connection.
1145* The target host must have the same or higher versions of the
1146 {pve} packages. (It *might* work the other way, but this is never guaranteed)
1147
1148Offline Migration
1149~~~~~~~~~~~~~~~~~
1150
c73c190f
DM
1151If you have local resources, you can still offline migrate your VMs,
1152as long as all disk are on storages, which are defined on both hosts.
1153Then the migration will copy the disk over the network to the target host.
1154
eeb87f95
DM
1155[[qm_copy_and_clone]]
1156Copies and Clones
1157-----------------
9e55c76d 1158
1ff5e4e8 1159[thumbnail="screenshot/gui-qemu-full-clone.png"]
9e55c76d
DM
1160
1161VM installation is usually done using an installation media (CD-ROM)
61018238 1162from the operating system vendor. Depending on the OS, this can be a
9e55c76d
DM
1163time consuming task one might want to avoid.
1164
1165An easy way to deploy many VMs of the same type is to copy an existing
1166VM. We use the term 'clone' for such copies, and distinguish between
1167'linked' and 'full' clones.
1168
1169Full Clone::
1170
1171The result of such copy is an independent VM. The
1172new VM does not share any storage resources with the original.
1173+
707e37a2 1174
9e55c76d
DM
1175It is possible to select a *Target Storage*, so one can use this to
1176migrate a VM to a totally different storage. You can also change the
1177disk image *Format* if the storage driver supports several formats.
1178+
707e37a2 1179
730fbca4 1180NOTE: A full clone needs to read and copy all VM image data. This is
9e55c76d 1181usually much slower than creating a linked clone.
707e37a2
DM
1182+
1183
1184Some storage types allows to copy a specific *Snapshot*, which
1185defaults to the 'current' VM data. This also means that the final copy
1186never includes any additional snapshots from the original VM.
1187
9e55c76d
DM
1188
1189Linked Clone::
1190
730fbca4 1191Modern storage drivers support a way to generate fast linked
9e55c76d
DM
1192clones. Such a clone is a writable copy whose initial contents are the
1193same as the original data. Creating a linked clone is nearly
1194instantaneous, and initially consumes no additional space.
1195+
707e37a2 1196
9e55c76d
DM
1197They are called 'linked' because the new image still refers to the
1198original. Unmodified data blocks are read from the original image, but
1199modification are written (and afterwards read) from a new
1200location. This technique is called 'Copy-on-write'.
1201+
707e37a2
DM
1202
1203This requires that the original volume is read-only. With {pve} one
1204can convert any VM into a read-only <<qm_templates, Template>>). Such
1205templates can later be used to create linked clones efficiently.
1206+
1207
730fbca4
OB
1208NOTE: You cannot delete an original template while linked clones
1209exist.
9e55c76d 1210+
707e37a2
DM
1211
1212It is not possible to change the *Target storage* for linked clones,
1213because this is a storage internal feature.
9e55c76d
DM
1214
1215
1216The *Target node* option allows you to create the new VM on a
1217different node. The only restriction is that the VM is on shared
1218storage, and that storage is also available on the target node.
1219
730fbca4 1220To avoid resource conflicts, all network interface MAC addresses get
9e55c76d
DM
1221randomized, and we generate a new 'UUID' for the VM BIOS (smbios1)
1222setting.
1223
1224
707e37a2
DM
1225[[qm_templates]]
1226Virtual Machine Templates
1227-------------------------
1228
1229One can convert a VM into a Template. Such templates are read-only,
1230and you can use them to create linked clones.
1231
1232NOTE: It is not possible to start templates, because this would modify
1233the disk images. If you want to change the template, create a linked
1234clone and modify that.
1235
319d5325
DC
1236VM Generation ID
1237----------------
1238
941ff8d3 1239{pve} supports Virtual Machine Generation ID ('vmgenid') footnote:[Official
effa4818
TL
1240'vmgenid' Specification
1241https://docs.microsoft.com/en-us/windows/desktop/hyperv_v2/virtual-machine-generation-identifier]
1242for virtual machines.
1243This can be used by the guest operating system to detect any event resulting
1244in a time shift event, for example, restoring a backup or a snapshot rollback.
319d5325 1245
effa4818
TL
1246When creating new VMs, a 'vmgenid' will be automatically generated and saved
1247in its configuration file.
319d5325 1248
effa4818
TL
1249To create and add a 'vmgenid' to an already existing VM one can pass the
1250special value `1' to let {pve} autogenerate one or manually set the 'UUID'
1251footnote:[Online GUID generator http://guid.one/] by using it as value,
1252e.g.:
319d5325 1253
effa4818
TL
1254----
1255 qm set VMID -vmgenid 1
1256 qm set VMID -vmgenid 00000000-0000-0000-0000-000000000000
1257----
319d5325 1258
cfd48f55
TL
1259NOTE: The initial addition of a 'vmgenid' device to an existing VM, may result
1260in the same effects as a change on snapshot rollback, backup restore, etc., has
1261as the VM can interpret this as generation change.
1262
effa4818
TL
1263In the rare case the 'vmgenid' mechanism is not wanted one can pass `0' for
1264its value on VM creation, or retroactively delete the property in the
1265configuration with:
319d5325 1266
effa4818
TL
1267----
1268 qm set VMID -delete vmgenid
1269----
319d5325 1270
effa4818
TL
1271The most prominent use case for 'vmgenid' are newer Microsoft Windows
1272operating systems, which use it to avoid problems in time sensitive or
cfd48f55
TL
1273replicate services (e.g., databases, domain controller
1274footnote:[https://docs.microsoft.com/en-us/windows-server/identity/ad-ds/get-started/virtual-dc/virtualized-domain-controller-architecture])
1275on snapshot rollback, backup restore or a whole VM clone operation.
319d5325 1276
c069256d
EK
1277Importing Virtual Machines and disk images
1278------------------------------------------
56368da8
EK
1279
1280A VM export from a foreign hypervisor takes usually the form of one or more disk
59552707 1281 images, with a configuration file describing the settings of the VM (RAM,
56368da8
EK
1282 number of cores). +
1283The disk images can be in the vmdk format, if the disks come from
59552707
DM
1284VMware or VirtualBox, or qcow2 if the disks come from a KVM hypervisor.
1285The most popular configuration format for VM exports is the OVF standard, but in
1286practice interoperation is limited because many settings are not implemented in
1287the standard itself, and hypervisors export the supplementary information
56368da8
EK
1288in non-standard extensions.
1289
1290Besides the problem of format, importing disk images from other hypervisors
1291may fail if the emulated hardware changes too much from one hypervisor to
1292another. Windows VMs are particularly concerned by this, as the OS is very
1293picky about any changes of hardware. This problem may be solved by
1294installing the MergeIDE.zip utility available from the Internet before exporting
1295and choosing a hard disk type of *IDE* before booting the imported Windows VM.
1296
59552707 1297Finally there is the question of paravirtualized drivers, which improve the
56368da8
EK
1298speed of the emulated system and are specific to the hypervisor.
1299GNU/Linux and other free Unix OSes have all the necessary drivers installed by
1300default and you can switch to the paravirtualized drivers right after importing
59552707 1301the VM. For Windows VMs, you need to install the Windows paravirtualized
56368da8
EK
1302drivers by yourself.
1303
1304GNU/Linux and other free Unix can usually be imported without hassle. Note
eb01c5cf 1305that we cannot guarantee a successful import/export of Windows VMs in all
56368da8
EK
1306cases due to the problems above.
1307
c069256d
EK
1308Step-by-step example of a Windows OVF import
1309~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~
56368da8 1310
59552707 1311Microsoft provides
c069256d 1312https://developer.microsoft.com/en-us/windows/downloads/virtual-machines/[Virtual Machines downloads]
144d5ede 1313 to get started with Windows development.We are going to use one of these
c069256d 1314to demonstrate the OVF import feature.
56368da8 1315
c069256d
EK
1316Download the Virtual Machine zip
1317^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^
56368da8 1318
144d5ede 1319After getting informed about the user agreement, choose the _Windows 10
c069256d 1320Enterprise (Evaluation - Build)_ for the VMware platform, and download the zip.
56368da8
EK
1321
1322Extract the disk image from the zip
1323^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^
1324
c069256d
EK
1325Using the `unzip` utility or any archiver of your choice, unpack the zip,
1326and copy via ssh/scp the ovf and vmdk files to your {pve} host.
56368da8 1327
c069256d
EK
1328Import the Virtual Machine
1329^^^^^^^^^^^^^^^^^^^^^^^^^^
56368da8 1330
c069256d
EK
1331This will create a new virtual machine, using cores, memory and
1332VM name as read from the OVF manifest, and import the disks to the +local-lvm+
1333 storage. You have to configure the network manually.
56368da8 1334
c069256d 1335 qm importovf 999 WinDev1709Eval.ovf local-lvm
56368da8 1336
c069256d 1337The VM is ready to be started.
56368da8 1338
c069256d
EK
1339Adding an external disk image to a Virtual Machine
1340~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~
56368da8 1341
144d5ede 1342You can also add an existing disk image to a VM, either coming from a
c069256d
EK
1343foreign hypervisor, or one that you created yourself.
1344
1345Suppose you created a Debian/Ubuntu disk image with the 'vmdebootstrap' tool:
1346
1347 vmdebootstrap --verbose \
67d59a35 1348 --size 10GiB --serial-console \
c069256d
EK
1349 --grub --no-extlinux \
1350 --package openssh-server \
1351 --package avahi-daemon \
1352 --package qemu-guest-agent \
1353 --hostname vm600 --enable-dhcp \
1354 --customize=./copy_pub_ssh.sh \
1355 --sparse --image vm600.raw
1356
1357You can now create a new target VM for this image.
1358
1359 qm create 600 --net0 virtio,bridge=vmbr0 --name vm600 --serial0 socket \
1360 --bootdisk scsi0 --scsihw virtio-scsi-pci --ostype l26
56368da8 1361
c069256d
EK
1362Add the disk image as +unused0+ to the VM, using the storage +pvedir+:
1363
1364 qm importdisk 600 vm600.raw pvedir
1365
1366Finally attach the unused disk to the SCSI controller of the VM:
1367
1368 qm set 600 --scsi0 pvedir:600/vm-600-disk-1.raw
1369
1370The VM is ready to be started.
707e37a2 1371
7eb69fd2 1372
16b4185a 1373ifndef::wiki[]
7eb69fd2 1374include::qm-cloud-init.adoc[]
16b4185a
DM
1375endif::wiki[]
1376
6e4c46c4
DC
1377ifndef::wiki[]
1378include::qm-pci-passthrough.adoc[]
1379endif::wiki[]
16b4185a 1380
c2c8eb89 1381Hookscripts
91f416b7 1382-----------
c2c8eb89
DC
1383
1384You can add a hook script to VMs with the config property `hookscript`.
1385
c961e4cc 1386 qm set 100 --hookscript local:snippets/hookscript.pl
c2c8eb89
DC
1387
1388It will be called during various phases of the guests lifetime.
1389For an example and documentation see the example script under
1390`/usr/share/pve-docs/examples/guest-example-hookscript.pl`.
7eb69fd2 1391
88a31964
DC
1392[[qm_hibernate]]
1393Hibernation
1394-----------
1395
1396You can suspend a VM to disk with the GUI option `Hibernate` or with
1397
1398 qm suspend ID --todisk
1399
1400That means that the current content of the memory will be saved onto disk
1401and the VM gets stopped. On the next start, the memory content will be
1402loaded and the VM can continue where it was left off.
1403
1404[[qm_vmstatestorage]]
1405.State storage selection
1406If no target storage for the memory is given, it will be automatically
1407chosen, the first of:
1408
14091. The storage `vmstatestorage` from the VM config.
14102. The first shared storage from any VM disk.
14113. The first non-shared storage from any VM disk.
14124. The storage `local` as a fallback.
1413
8c1189b6 1414Managing Virtual Machines with `qm`
dd042288 1415------------------------------------
f69cfd23 1416
dd042288 1417qm is the tool to manage Qemu/Kvm virtual machines on {pve}. You can
f69cfd23
DM
1418create and destroy virtual machines, and control execution
1419(start/stop/suspend/resume). Besides that, you can use qm to set
1420parameters in the associated config file. It is also possible to
1421create and delete virtual disks.
1422
dd042288
EK
1423CLI Usage Examples
1424~~~~~~~~~~~~~~~~~~
1425
b01b1f2c
EK
1426Using an iso file uploaded on the 'local' storage, create a VM
1427with a 4 GB IDE disk on the 'local-lvm' storage
dd042288 1428
b01b1f2c 1429 qm create 300 -ide0 local-lvm:4 -net0 e1000 -cdrom local:iso/proxmox-mailgateway_2.1.iso
dd042288
EK
1430
1431Start the new VM
1432
1433 qm start 300
1434
1435Send a shutdown request, then wait until the VM is stopped.
1436
1437 qm shutdown 300 && qm wait 300
1438
1439Same as above, but only wait for 40 seconds.
1440
1441 qm shutdown 300 && qm wait 300 -timeout 40
1442
87927c65
DJ
1443Destroying a VM always removes it from Access Control Lists and it always
1444removes the firewall configuration of the VM. You have to activate
1445'--purge', if you want to additionally remove the VM from replication jobs,
1446backup jobs and HA resource configurations.
1447
1448 qm destroy 300 --purge
1449
1450
f0a8ab95
DM
1451
1452[[qm_configuration]]
f69cfd23
DM
1453Configuration
1454-------------
1455
f0a8ab95
DM
1456VM configuration files are stored inside the Proxmox cluster file
1457system, and can be accessed at `/etc/pve/qemu-server/<VMID>.conf`.
1458Like other files stored inside `/etc/pve/`, they get automatically
1459replicated to all other cluster nodes.
f69cfd23 1460
f0a8ab95
DM
1461NOTE: VMIDs < 100 are reserved for internal purposes, and VMIDs need to be
1462unique cluster wide.
1463
1464.Example VM Configuration
1465----
777cf894 1466boot: order=virtio0;net0
f0a8ab95
DM
1467cores: 1
1468sockets: 1
1469memory: 512
1470name: webmail
1471ostype: l26
f0a8ab95
DM
1472net0: e1000=EE:D2:28:5F:B6:3E,bridge=vmbr0
1473virtio0: local:vm-100-disk-1,size=32G
1474----
1475
1476Those configuration files are simple text files, and you can edit them
1477using a normal text editor (`vi`, `nano`, ...). This is sometimes
1478useful to do small corrections, but keep in mind that you need to
1479restart the VM to apply such changes.
1480
1481For that reason, it is usually better to use the `qm` command to
1482generate and modify those files, or do the whole thing using the GUI.
1483Our toolkit is smart enough to instantaneously apply most changes to
1484running VM. This feature is called "hot plug", and there is no
1485need to restart the VM in that case.
1486
1487
1488File Format
1489~~~~~~~~~~~
1490
1491VM configuration files use a simple colon separated key/value
1492format. Each line has the following format:
1493
1494-----
1495# this is a comment
1496OPTION: value
1497-----
1498
1499Blank lines in those files are ignored, and lines starting with a `#`
1500character are treated as comments and are also ignored.
1501
1502
1503[[qm_snapshots]]
1504Snapshots
1505~~~~~~~~~
1506
1507When you create a snapshot, `qm` stores the configuration at snapshot
1508time into a separate snapshot section within the same configuration
1509file. For example, after creating a snapshot called ``testsnapshot'',
1510your configuration file will look like this:
1511
1512.VM configuration with snapshot
1513----
1514memory: 512
1515swap: 512
1516parent: testsnaphot
1517...
1518
1519[testsnaphot]
1520memory: 512
1521swap: 512
1522snaptime: 1457170803
1523...
1524----
1525
1526There are a few snapshot related properties like `parent` and
1527`snaptime`. The `parent` property is used to store the parent/child
1528relationship between snapshots. `snaptime` is the snapshot creation
1529time stamp (Unix epoch).
f69cfd23 1530
88a31964
DC
1531You can optionally save the memory of a running VM with the option `vmstate`.
1532For details about how the target storage gets chosen for the VM state, see
1533xref:qm_vmstatestorage[State storage selection] in the chapter
1534xref:qm_hibernate[Hibernation].
f69cfd23 1535
80c0adcb 1536[[qm_options]]
a7f36905
DM
1537Options
1538~~~~~~~
1539
1540include::qm.conf.5-opts.adoc[]
1541
f69cfd23
DM
1542
1543Locks
1544-----
1545
0bcc62dd
DM
1546Online migrations, snapshots and backups (`vzdump`) set a lock to
1547prevent incompatible concurrent actions on the affected VMs. Sometimes
1548you need to remove such a lock manually (e.g., after a power failure).
f69cfd23
DM
1549
1550 qm unlock <vmid>
1551
0bcc62dd
DM
1552CAUTION: Only do that if you are sure the action which set the lock is
1553no longer running.
1554
f69cfd23 1555
16b4185a
DM
1556ifdef::wiki[]
1557
1558See Also
1559~~~~~~~~
1560
1561* link:/wiki/Cloud-Init_Support[Cloud-Init Support]
1562
1563endif::wiki[]
1564
1565
f69cfd23 1566ifdef::manvolnum[]
704f19fb
DM
1567
1568Files
1569------
1570
1571`/etc/pve/qemu-server/<VMID>.conf`::
1572
1573Configuration file for the VM '<VMID>'.
1574
1575
f69cfd23
DM
1576include::pve-copyright.adoc[]
1577endif::manvolnum[]